Described is a system (and method) for providing a flexible decision system for autonomous driving. The system may include a framework that allows a decision system to switch between a deliberate rule-based decision framework and an intuitive machine-learning model-based decision framework. The system may invoke the appropriate framework (or subsystem) based on a particular set of driving conditions or environment. Accordingly, the system described herein may provide an efficient and adaptable decision system for autonomous driving.
Legal claims defining the scope of protection, as filed with the USPTO.
1. A computer-implemented method for providing autonomous driving control for a vehicle, comprising: receiving one or more types of data during operation of the vehicle, wherein the one or more types of data include at least perception data of a driving environment surrounding the vehicle; and providing autonomous driving control, by the vehicle, wherein the driving control switches between a deliberation mode and an intuitive mode based on an analysis of the one or more types of data, wherein the deliberation mode provides the autonomous driving control based on a set of predefined driving rules, and the intuitive mode provides the autonomous driving control based on a machine-learning model, and wherein the autonomous driving control includes a selected route for the vehicle as well as one or more driving parameters, wherein output from the deliberation mode is provided to the machine-learning model in the intuitive mode as training data, including controlling the vehicle in the intuitive mode based on the machine-learning model including a deep neural network (DNN) as a default mode to reduce computational resources, and switching from the intuitive mode based on the machine-learning model including the DNN to the deliberation mode based on the set of predefined driving rules including traffic rules, route rules, and ride comfort rules, in response to the analysis of the one or more types of data determining a complexity threshold of the driving environment is met, wherein the complexity threshold is determined based on the number of perceived vehicle or pedestrian objects within the driving environment surrounding the vehicle.
2. The method of claim 1 , wherein the driving control switches to the intuitive mode in response to the analysis of the one or more types of data determining a familiarity threshold of the driving environment is satisfied.
3. The method of claim 1 , wherein the one or more types of data further include map and real-time traffic information, and location information of the vehicle.
4. The method of claim 1 , wherein the one or more types of data further include location information of the vehicle.
5. A non-transitory machine-readable medium having instructions stored therein, which when executed by a processor, cause the processor to perform operations for providing autonomous driving control for a vehicle, the operations comprising: receiving one or more types of data during operation of the vehicle, wherein the one or more types of data include at least perception data of a driving environment surrounding the vehicle; and providing autonomous driving control, by the vehicle, wherein the driving control switches between rule-based decision making and model-based decision making based on an analysis of the one or more types of data, wherein when the rule-based decision making is invoked, the autonomous driving control is based on a set of predefined driving rules, and when the model-based decision making is invoked, the autonomous driving control is based on a machine-learning model, and wherein the autonomous driving control includes a selected route for the vehicle as well as one or more driving parameters, wherein output from the deliberation mode is provided to the machine-learning model in the intuitive mode as training data, including controlling the vehicle in the intuitive mode based on the machine-learning model including a deep neural network (DNN) as a default mode to reduce computational resources, and switching from the intuitive mode based on the machine-learning model including the DNN to the deliberation mode based on the set of predefined driving rules including traffic rules, route rules, and ride comfort rules, in response to the analysis of the one or more types of data determining a complexity threshold of the driving environment is met, wherein the complexity threshold is determined based on the number of perceived vehicle or pedestrian objects within the driving environment surrounding the vehicle.
6. The medium of claim 5 , wherein the analysis of the one or more types of data includes determining a familiarity threshold of the driving environment.
7. The medium of claim 5 , wherein the one or more types of data further include map and real-time traffic information.
8. The medium of claim 5 , wherein the one or more types of data further include location information of the vehicle.
9. A data processing system for providing autonomous driving control for a vehicle, comprising: a processor; and a memory coupled to the processor to store instructions, which when executed by the processor, cause the processor to perform operations, the operations including receiving one or more types of data during operation of the vehicle, wherein the one or more types of data include at least perception data of a driving environment surrounding the vehicle; and providing autonomous driving control, by the vehicle, using a decision system, wherein the decision system switches between a deliberation subsystem and an intuitive subsystem based on an analysis of the one or more types of data, wherein the deliberation subsystem provides the autonomous driving control based on a set of predefined driving rules, and the intuitive subsystem provides the autonomous driving control based on a machine-learning model, and wherein the autonomous driving control includes a selected route for the vehicle as well as one or more driving parameters, wherein output from the deliberation mode is provided to the machine-learning model in the intuitive mode as training data, including controlling the vehicle in the intuitive mode based on the machine-learning model including a deep neural network (DNN) as a default mode to reduce computational resources, and switching from the intuitive mode based on the machine-learning model including the DNN to the deliberation mode based on the set of predefined driving rules including traffic rules, route rules, and ride comfort rules, in response to the analysis of the one or more types of data determining a complexity threshold of the driving environment is met, wherein the complexity threshold is determined based on the number of perceived vehicle or pedestrian objects within the driving environment surrounding the vehicle.
10. The system of claim 9 , wherein the decision system switches to the intuitive subsystem in response to the analysis of the one or more types of data determining a familiarity threshold of the driving environment is satisfied.
11. The system of claim 9 , wherein the one or more types of data further include map, real-time traffic information, and location information of the vehicle.
Cooperative Patent Classification codes for this invention. Click any code to explore related patents in that topic.
June 2, 2017
October 27, 2020
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.